Naturally, Professor Dennett can't afford to read nor certainly respond to anonymous emails from his fandom, but at least I got a very nice canned response explaining that.
Still, even though it's as dense and therefore time-consuming to read as everything I write, I thought I'd post it here, just for the record.
Of course, since writing it, my own thinking has evolved, and Professor Bem was kind enough to remind me that he has refuted his refuters, and more importantly, that he did, in fact, use both a pseudo random number generator and a true random number generator in his experiments.
So, while I would drop my proposed "experiment" I still maintain that all that can be tested for is the reasonable limits for a definition of mind (its boundary) and not pre-cognition per se. Or in other words, if mind cannot be dis-implicated from matter (a major burden of Dennett's work) then the imposition of time's arrow, which is difficult to accomplish *in* the mind, is difficult to distinguish from fictional or creative narrative.
It is, of course, my burden to de-center notions of narrtive true-ing processes more generally, and I carry on:
I've just finished reading Consciousness Explained and then familiarizing myself with your work more generally. In brief, I find the writing brilliant and the overall case cogent and convincing. Since it's far too late for me to take up the study of consciousness to the level of your accomplishment, I feel entirely comfortable saying that for my purposes you have indeed explained consciousness. (I'll still read on through what you've written since!)
Here's why I write: As it happens, I read your book in light of the soon to be published work of Professor Bem regarding Psi and possibly pre-cognition of which I'm sure you are aware. I also read a cogent critique from the Bayesian side of statistics which relieves me, at least, of the need to worry too much about pre-cognition as Bem's work purports to discover it. Were there pre-cognition it would, of course, fall within the margin of error which Bayesian statistics provides as corrective to non-Bayesian methods. That's almost definitional. Accounting for the priors is hard to distinguish from accounting for frame of mind.
But since your approach to consciousness defies attempts to "locate" it, say in a brain, and since therefore there can be no meaning to simultaneity among drafted cognitions in a single mind - as you demonstrate convincingly in your book - therefore time's arrow has no place "in" the mind. A properly time-sequenced narrative must emerge for sense to prevail, but that is a separate matter from the order of events perceived, conceived, cognized or re-cognized.
You also challenge, at least implicitly, the bordering of the mind by the limits of its physical substrate (the brain, for instance) among other places by your suggestion that a single mind may be thought to be distributed, as in certain twins who inhabit a coherent biography.
So first of all, if there is no singular locus for any thought inside the brain, any cognition there must in effect allow for pre-cognition. Indeed it's not a problem worth bothering about, since even in principle there's no way to measure it, again as I believe you amply demonstrate in this book. At the very least any thought remains subject to subsequent cognition for so long as it might be held in abeyance by some sense that there will be more to assimilate before a completed thought is uttered.
Consciousness, in other words, may have more to do with sense of completion, rather more like the ballistic act of throwing a ball than like hitting a target. You initiate and correct along the way until release.
But more interestingly to me, if the mind cannot be located entirely "inside" the brain, then all those peripheral happenings which impinge on thought, of course including chance or random events, could be thought to condition moment of release according to their various potentials and probabilities (as variously perceived or conceived) regardless of their literal time-sequencing.
In Bem's experiment, the button-clickers have some anticipation of porn - they have a motive to click, not unlike someone playing Jeopardy.
Pre-cognition, however, is the wrong thing to test for. Rather, there should be some test along the lines of how Bell's Theorem is tested in physics, to demonstrate the impossibility to disprove spatial separation of cognitive inputs (technically indistinguishable from pre-cognition) to within the margin of error between Bayesian and non-Bayesian statistical models.
My suggestion would be to redo the Bem experiment but using the pseudo-random number generated by the bounded system of the computer, rather than the "true" random number generated in the same "cosmos" within which the subject's mind is choosing.
After all, it is damned impossible to determine that the coin is a fair coin after a long run of heads, other than by invoking "prior" knowledge. But as with voting machines which preserve a physical record, as least with a pseudo random number, you have a trace to compare against.
Of course it is my sense that, as with physical reality, there is some dimension beyond which certainty is not only impossible for practical reasons, but impossible in principle. I suspect that this prospect unsettles you no more than it does me. I'd love to learn of a better conceived experiment than the one I propose, but I can't think one up right now - I know it's out there!
Of course I have plenty more to say on the subject, but I have used enough of your time (pure wishful thinking on my part).